AgentChat

AgentChat is a modern intelligent dialogue system that orchestrates multiple AI models, tools, and knowledge sources. It integrates MCP servers to enable Model Context Protocol-based services, allowing you to upload and manage custom MCP services alongside built-in capabilities such as knowledge retrieval, tool invocation, and multi-agent workflows. Once the server is running, you can use the MCP-enabled tools to perform tasks like querying knowledge bases, calling external tools (e.g., weather, document parsing, search), and coordinating multiple agents to solve complex tasks. The frontend provides an intuitive dashboard for managing agents, tools, and workflows, while the backend handles streaming responses and asynchronous task execution for real-time interactions.

Prerequisites:

• Python 3.12+ installed on your system
• Git installed
• Optional: Docker if you prefer containerized deployment

1. Clone the repository

• git clone https://github.com/shy2593666979/agentchat.git
• cd agentchat

2. Create a virtual environment (recommended)

• python -m venv venv
• source venv/bin/activate # on macOS/Linux
• .\venv\Scripts\activate # on Windows

3. Install dependencies

• pip install -U pip setuptools wheel
• pip install -r requirements.txt

4. Configure environment variables

• Copy .env.example to .env
• Edit .env with appropriate values (DATABASE_URL, REDIS_URL, OPENAI_API_KEY, etc.)

5. Run the server

• python -m agentchat

6. Optional: run with Docker

• Ensure Docker is running
• Build: docker build -t agentchat:latest .
• Run: docker run -i --env-file .env -p 8000:8000 agentchat:latest

7. Verify

• Access http://localhost:8000 (or the configured port) and follow the documentation for endpoints and MCP configuration.

Tips and common issues:

• Ensure your OPENAI_API_KEY (or other model keys) are set if you rely on external AI services.
• When using MCP, make sure the MCP endpoints you upload are compatible with the server's expected protocol version. Check migration notes if upgrading from older versions.
• For multi-agent workflows, define clear goals and constraints to avoid miscoordination.
• If you encounter startup errors related to missing dependencies, re-create the virtual environment and reinstall requirements.
• Use the environment variable APP_ENV to toggle between development and production modes; enable logging and monitoring in production for better observability.